Files
BaaRai 28e6618c71 feat: surface MCP server connect health to operators and agents
Every connect attempt (discovery, refresh, and the per-agent-run wiring) now
records its outcome on the server row: last_connect_at/ok/err, written through
a change-deduped update so per-run wiring does not amplify writes. The MCP page
shows a health chip per card (ok with time, or the failure reason), and a new
read-only mcp_status tool (bound to auto, worker and pentest) lets a worker
notice that a server it expected is down and say so instead of silently
degrading while the cached tool count keeps the card looking healthy.

Closes #38
2026-10-11 22:50:00 +02:00

94 lines
3.0 KiB
Go

package server
import (
"context"
"fmt"
"log"
"time"
"github.com/Autumn-27/artex/db"
"github.com/Autumn-27/artex/mcphttp"
"github.com/Autumn-27/norma/mcp"
actool "github.com/Autumn-27/norma/tool"
)
// mcpClient is the shared surface of a connected MCP server (stdio, Streamable HTTP,
// or legacy SSE),
// so tools/list and cleanup are handled uniformly regardless of transport.
type mcpClient interface {
Tools(context.Context) ([]actool.CoreTool, error)
Close() error
}
// connectMCP dials one MCP server per its transport. Callers must Close the client.
func connectMCP(ctx context.Context, m *db.MCPServer) (mcpClient, error) {
switch m.Transport {
case "stdio":
if m.Command == "" {
return nil, fmt.Errorf("stdio transport is missing a command")
}
return mcp.NewStdioClient(ctx, m.Name, m.Command, jsonStrMap(m.Env), jsonStrSlice(m.Args)...)
case "http":
if m.URL == "" {
return nil, fmt.Errorf("http transport is missing a URL")
}
// env map doubles as HTTP headers (e.g. Authorization).
return mcphttp.New(ctx, m.Name, m.URL, jsonStrMap(m.Env), m.Insecure, m.TimeoutSeconds)
case "sse":
if m.URL == "" {
return nil, fmt.Errorf("sse transport is missing a URL")
}
return mcphttp.NewSSE(ctx, m.Name, m.URL, jsonStrMap(m.Env), m.Insecure)
default:
return nil, fmt.Errorf("unknown transport %q", m.Transport)
}
}
// discoverAndCacheMCP connects to one MCP, lists its tools, and persists the tool
// names to mcp_tools_cache so the UI shows them without a live connection.
func (s *Server) discoverAndCacheMCP(ctx context.Context, m *db.MCPServer) error {
cl, err := connectMCP(ctx, m)
if err != nil {
_ = s.m.pg.UpdateMCPHealth(m.ID, false, err.Error())
return err
}
defer cl.Close()
ts, err := cl.Tools(ctx)
if err != nil {
_ = s.m.pg.UpdateMCPHealth(m.ID, false, err.Error())
return err
}
_ = s.m.pg.UpdateMCPHealth(m.ID, true, "")
tools := make([]db.MCPTool, 0, len(ts))
for _, t := range ts {
tools = append(tools, db.MCPTool{Name: t.Name(), Description: t.Description()})
}
if err := s.m.pg.SaveMCPTools(m.ID, tools); err != nil {
return err
}
log.Printf("[mcp] %s discovered %d tools and cached them", m.Name, len(tools))
return nil
}
// discoverEmptyMCPsOnStartup fills the tool cache for any enabled MCP that has none
// yet (notably the seeded browser MCP on first run). Runs sequentially in one
// goroutine so we never spawn many stdio servers (npx) at once, and never blocks
// startup. Best-effort: a failure leaves the cache empty to retry next start.
func (s *Server) discoverEmptyMCPsOnStartup() {
servers, err := s.m.pg.ListMCP()
if err != nil {
log.Printf("[mcp] startup auto-discovery: failed to read the list: %v", err)
return
}
for _, m := range servers {
if !m.Enabled || len(m.Tools) > 0 {
continue
}
ctx, cancel := context.WithTimeout(s.ctx, 90*time.Second)
if err := s.discoverAndCacheMCP(ctx, m); err != nil {
log.Printf("[mcp] startup auto-discovery of %s failed: %v", m.Name, err)
}
cancel()
}
}